Edexcel Psychology for A Level Year 2: Student Book

Chapter 1 Issues and debates Practical issues in the design and implementation of research Validity Validity versus ethics As we have seen, ethical issues arise when researchers do not reveal their true aims or are ‘economical’ with the truth about their procedures. For researchers, participants’ naivety reduces the risk of demand characteristics. For example, in memory studies, keeping participants uninformed prevents them making a specific effort to remember particular stimuli. Social desirability It is important to design questionnaires and interviews to limit a participant’s wish to ‘look good’. For example, in research on prejudice, participants may lie about their beliefs or behaviours to hide their prejudiced, and therefore unacceptable, views. Similarly in clinical research, participants may be reluctant to reveal symptoms if they perceive them to be undesirable. This results in under-representation of mental illness or of particular symptoms within disorders. Measurement In order to produce useful results, researchers must measure variables in meaningful (valid) ways. In the psychodynamic approach, measures of the unconscious are highly subjective and their credibility is questionable. If an abstract concept such as the ‘id’ cannot be directly observed it is difficult to verify that it exists, so a valid measure of id strength is hard to develop. Measures of brain function present a different problem. The measures themselves are valid – a more active PET scan means there is more activity in that area of the brain. The difficulty is in making valid interpretations about what such differences or changes in brain activity mean . In clinical psychology it is difficult to be objective about symptoms. For example, in schizophrenia, hallucinations are personal experiences so they are hard to measure in a way that is independent of the individual’s viewpoint. Controls By using controls in an experiment, the researcher minimises the influence of confounding variables. For example, in experiments on operant conditioning using Skinner boxes (Y1 page 134), the mechanised environment controls what a rat sees, hears and feels – such as lights, clicks and electric shocks. Furthermore, to ensure they are hungry (so they will press the bar for food), they are often kept at 80% of their normal weight. Ecological validity The setting and task of a study should allow the results to apply to other situations. This can be hard to achieve. For example, in order to achieve good controls and therefore high internal validity, studies in cognitive psychology often use word lists or nonsense syllables. Memory in the real world, however, rarely demands the recall of meaningless material so the tests used in research cannot reflect the reality of day-to-day remembering as this can use cues and links between information to avoid forgetting. In field experiments, the setting/task may have greater ecological validity but controls are harder to implement than in laboratory experiments. In areas such as social psychology both types of study are used so results can be compared. Reliability Reliability and standardisation Standardisation directly increases reliability by keeping procedures the same between participants. Standardised instructions are used in laboratory studies, such as on memory. However, this is harder to achieve in field experiments such as in social psychology. Reliability over time The test-retest measure of reliability detects whether measurement in studies, or diagnosis in clinical psychology, is consistent from one occasion to the next. Only by having reliable measures can we be sure than any changes, such as improvement or worsening of clinical symptoms over time, are due to the treatment being given. Reliability between researchers In research on mental illness, one way to collect data might be using an interview. The reliability of the interview could be assessed by having two interviewers and comparing the answers they collect (inter-rater reliability). Alternatively behaviour of participants might be observed and then reliability can be checked by having two observers and comparing the record of their observations (inter-observer reliability). One reason for low reliability is the collection of qualitative data (in interviews and observations). In general, qualitative data requires interpretation so has the potential to lead to differences between researchers. Quantitative data on the other hand is more objective and thus more reliable. Generalisability Generalisability relates to ecological validity, but other issues also affect how well findings apply to the target population. The sample Wider samples typically have greater generalisability but this is not always the main priority. In clinical psychology, for example, narrow samples may be selected to focus on people with a particular symptom, or to avoid using people with co-morbid conditions. Generalising from animals Many areas other than learning theories (such as clinical psychology) use animal research. The complexity of human brains, cognition, emotions and our social interactions can mean that such research lacks relevance to people. Participants just want to offer a helping hand – but that means they try to guess the aims of the study and shape their behaviour. Result = reduced validity. Applying your knowledge An example of a synoptic exam question linked to practical issues: Assess whether solving practical issues is the only barrier to good research. (20) The paragraphs below illustrate how you might answer this: High validity is an important consideration when designing research but to achieve this often means that other issues are compromised. Validity in studies of obedience relies on deceiving participants. This is essential to avoid demand characteristics but means informed consent cannot be given. So the best research sometimes has to be unethical. (52 words) Reliability matters to good research and can depend on the type of data. Clinical psychology often uses qualitative data. This may be valid, e.g. descriptions of symptoms by clients, but also subjective so interpretation can lead to differences over time or between clinicians. Therefore, good reliability is hard to achieve. (50 words) 18